DataFrame常用功能及技巧

1.*滑动窗口-rolling*
2.*浮点数保留3位小数*

1、滑动窗口-rolling

官方文档: Window 接口

pandas.api.typing.Rolling instances are returned by .rolling calls:

pandas.DataFrame.rolling()
pandas.Series.rolling().

pandas.api.typing.Expanding instances are returned by .expanding calls:     pandas.DataFrame.expanding()
andas.Series.expanding().

pandas.api.typing.ExponentialMovingWindow instances are returned by .ewm calls:   pandas.DataFrame.ewm()
pandas.Series.ewm().

以下是主要的函数及说明 - Rolling window functions

Rolling.count([numeric_only])                   Calculate the rolling count of non NaN observations.

Rolling.sum([numeric_only, engine, ...])        Calculate the rolling sum.

Rolling.mean([numeric_only, engine, ...])       Calculate the rolling mean.

Rolling.median([numeric_only, engine, ...])     Calculate the rolling median.

Rolling.var([ddof, numeric_only, engine, ...])  Calculate the rolling variance.

Rolling.std([ddof, numeric_only, engine, ...])  Calculate the rolling standard deviation.

Rolling.min([numeric_only, engine, ...])        Calculate the rolling minimum.

Rolling.max([numeric_only, engine, ...])        Calculate the rolling maximum.

Rolling.corr([other, pairwise, ddof, ...])      Calculate the rolling correlation.

Rolling.cov([other, pairwise, ddof, ...])       Calculate the rolling sample covariance.

Rolling.skew([numeric_only])                    Calculate the rolling unbiased skewness.

Rolling.kurt([numeric_only])                    Calculate the rolling Fisher's definition of kurtosis without bias.

Rolling.apply(func[, raw, engine, ...])         Calculate the rolling custom aggregation function.

Rolling.aggregate(func, \*args, \*\*kwargs)     Aggregate using one or more operations over the specified axis.

Rolling.quantile(q[, interpolation, ...])       Calculate the rolling quantile.

Rolling.sem([ddof, numeric_only])               Calculate the rolling standard error of mean.

Rolling.rank([method, ascending, pct, ...])     Calculate the rolling rank.

Weighted window functions

Window.mean([numeric_only])             Calculate the rolling weighted window mean.

Window.sum([numeric_only])              Calculate the rolling weighted window sum.

Window.var([ddof, numeric_only])        Calculate the rolling weighted window variance.

Window.std([ddof, numeric_only])        Calculate the rolling weighted window standard deviation.

2、浮点数保留3位小数

1）某一列的浮点数保留3位小数

# 保留3位小数
df['A'] = df['A'].round(3)

2）保存csv文件时，浮点数保留3位小数

dfData.to_csv('outfile.csv', index=False, encoding='gbk',float_format='%.3f')

3）使用 set_option() 方法设置小数点精度

dfData = pd.DataFrame()
print"DataFrame ...\n",dataFrame

# 设置 pd 使用小数精度
pd.set_option('display.precision', 2)

print"\n更新后带有小数点的DataFrame...\n", dataFrame